Poul-Henning,
it was the KS representative on our meeting, who expected an
updated/improved specification from the re-validation process, as the
new prototypes obviously performed better than the existing 3458As, in
some (undisclosed) parameters.
The specification origins from the earliest version of 1988. It was
never being updated, even after HP changed the 40k reference resistor
from a hermetical ~ S102C technology to a VHP101 from about 1995
onwards, which in reality performs much better than the original
resistor at 1ppm/K and 10ppm/year.
The VHP101 is an oil-filled, T.C. compensated type, who might have
~0.3ppm/K and 2ppm/6 years drift only, even if Vishay generally makes
very generous specifications.
Already hp or at least KS might have challenged / tested this during
their recent validation
They also might have improved the resistor network for the different Ohm
ranges for better T.C., as well as the mediocre DCI shunts.
Same goes for the DCV specification, they could have simply validated a
metrology grade specification, i.e. operating temperature from 18 to
38°C, and validating a LTZ1000 reference at 55..65°C with <2ppm/year,
<0.05ppm/K stability, with only small changes in design: 12 or 12.5k
instead of 15k, use of non-A version, T.C. trimming of the LTZ circuit,
and a proper thermally shielded case.
They only would have had to make a new validation of such a metrology
grade specification.
The BOM and monitoring cost would be lower, as yield would be higher
(near 100%), and they may have demanded even more money for that.
Maybe they could have modified the A/D converter firmware a bit, to
compensate the T.C. of the ADC resistor network by implementing an 'ACAL
converter' phase from time to time (measuring briefly the network gain
=> CAL? 72). That would also solve the U180 drift problem.
I also don't like proposals or assumptions, that KS should or would
design a completely new and improved 3458, I really think that they
suffer from brain-drain to do so.
But with those few small modifications, they really could have improved
the specification greatly to attack FLUKEs new Reference DMMs.
Frank
--------
In message <[email protected]>, Frank Stellmach w
rites:
Therefore, all specifications and manuals are identical.. a pity..
it was hoped for, that the specifications would be improved.
How or what did you expect them to improve ?
Improving the long term stability of the voltage and ohm references
would be nice, but either one would increase the price by 10% or so,
because we would be into "buy ten throw 9 away" territory.
Anything else which could deliver relevant improvements gets to be
complicated and expensive:
Battery power + optical interface could *possibly* reduce the noise
a little bit.
A custom input connector could *possibly* reduce thermal sensitivity.
External clock input *could* allow you to do synchronized sampling.
And yeah, you could change the interface to be SCPI, breaking all
softwre ever written for that model...
So all in all: I would hate to be pointed at and told "make a better one".
_______________________________________________
volt-nuts mailing list -- [email protected]
To unsubscribe, go to
http://lists.febo.com/mailman/listinfo/volt-nuts_lists.febo.com
and follow the instructions there.
